Vinci targets faster chip design and simulation with physics-driven AI

Simulation plays a crucial role in hardware engineering by helping identify issues in product designs and ensuring proper deployment. However, running these simulations often required specialized skills and significant time. To address this challenge, Vinci4D Inc., a Silicon Valley-based startup, has developed a software platform that drastically speeds up the simulation process.

By utilizing physics-driven AI, Vinci claims its tool is capable of running simulations a day after design changes instead of the traditional two-week turnaround. This accelerated simulation capability reduces product development cycles and enables faster shipping of improved products. Hardik Kabaria, Vinci's co-founder and CEO, explained that the company trained an AI model on physics equations, packaging it within software for seamless deployment.

This allows engineers to leverage Vinci's technology without requiring specialized skills. Kabaria highlighted that the goal is to synchronize the speed of physics simulation with that of AI inference, enabling customers to design optimal hardware. Vinci's technology also interfaces with advanced manufacturing requirements, infrastructure reliability, and long-term operational performance.

Chipmakers are increasingly focusing on the expertise and infrastructure needed to support complex manufacturing environments. Vinci's collaboration with Advanced Micro Devices (AMD) exemplifies this alignment, as AMD's MI355x hardware can be used for AI inference of Vinci's software across various stages of semiconductor production.

The company recently secured a $250 million funding round at a $1.5 billion valuation, with investors including Advent, Temasek, Xora, and AMD Ventures. Kabaria emphasized that success in AI and simulation tools will require collaboration across the entire hardware ecosystem.